Multi-user multiple-input multiple-output (MU-MIMO) transmission is becoming a new system technique to enable high system capacity in both the upcoming IEEE 802.11ac and the LTE (long-term evolution) standards. As compared to single-user MIMO (SU-MIMO), MU-MIMO has several key advantages. First, MU-MIMO allows for a direct gain in multiple access system capacity proportional to the number of access point antennas. Second, MU-MIMO allows the higher degree spatial multiplexing gain to be obtained without the need for higher number of antennas at the mobile stations by keeping the intelligence and cost at the access point. Third, MU-MIMO appears immune to most propagation limitations plaguing SU-MIMO communications because multiuser diversity can be extracted even in a simple line of sight (LOS) propagation environment. As a result, the LOS propagation, which causes degradation in single user spatial multiplexing schemes, is no longer a problem in the multiuser setting.
In contrast to the SU-MIMO transmission, where the mobile station receivers are equipped with sufficient number of antennas (equal to or greater than the number of spatial streams) and the capability of the signal processing to estimate the channel and to separate the spatial streams, it is crucial in a MU-MIMO transmission for the access points or routers to bear the most of the burden in the signal processing and hardware complexity to allow for simpler mobile station implementation. To achieve this aim, the access point or router should apply transmit beamforming (precoding), computed from channel information acquired in the MU-MIMO downlink channel sounding and feedback to achieve an orthogonal (or near-orthogonal) transmission of multiple streams to multiple users, i.e., eliminating (or reducing) the amount of mutual interference between the transmission to multiple mobile stations. Under this condition, each mobile station only receives the spatial stream(s) intended for itself and not the interference from the spatial stream(s) intended for other mobile stations. With reduced number of spatial streams directed toward individual mobile stations, all mobile stations only need to be equipped with sufficient number of antennas for processing the spatial streams intended for itself and not worrying about eliminating the interference from other spatial streams.
FIG. 1 (Prior Art) illustrates a typical SU-MIMO and MU-MIMO process in a wireless communication system. For SU-MIMO, the receiver spatial processing occurs during the long training fields (LTFs) in the preamble before the arrival of the data payload. The receiver spatial processing is performed on a packet-by-packet basis. If the receiver spatial processing is not done correctly, for example due to interference, only that packet (e.g., packet 11 in FIG. 1) is affected and the erroneous transmission can be corrected by acknowledgement and re-transmission protocol as long as interference is no longer present in subsequent re-transmission.
For MU-MIMO, on the other hand, the channel knowledge and antenna weights at the transmitter are not updated frequently. The process of channel sounding and feedback add a significant overhead to the system processing and it is typically performed at intervals comparable to the channel coherent time. Note that this sounding and feedback process for MU-MIMO is identical for that for transmit beamforming. The same discussion presented here would also apply to transmit beamforming. If the incorrect channel knowledge, e.g., due to received sounding signal corrupted by interference, is used at the transmitter, the sequence of frame exchanges based on the same transmit antenna weights are affected and the resultant communication errors are not correctible through the acknowledgement and re-transmission protocol. In the example of FIG. 1, when sounding and feedback 21 is corrupted by interference, the sequence of MU-MIMO frame exchanges 22, 23 . . . are affected. Additionally, since the transmit beamforming weight is computed from aggregate channel measurements from multiple receivers, one incorrect feedback may produce a corrupted transmit (precoding) weights for all devices involved in the sounding and feedback process. Thus, it is critical to ensure the integrity and/or quality of the channel information obtained from the sounding process.
Currently, there is no mechanism or protocol in the 802.11ac system to allow fast recovery from situation that sounding process is corrupted. Since wideband channel bandwidths are proposed in the 802.11ac, the sounding process is more susceptible to interferences, especially to sub-channel interferences. Although there is channel protection mechanism such as RTS and CTS protocol in the IEEE802.11 standards for improving the channel integrity and/or quality of the channel sounding and feedback process, such protection mechanism adds to overhead of the process and is typically not employed.